Gaussian Graphical Models Identified Food Intake Networks among Iranian Women with and without Breast Cancer: A Case-Control Study

Abstract

Background: Dietary patterns may be an important predictor of breast cancer risk. However, they cannot completely explain the pairwise correlations among foods. The purpose of this study is to compare food intake networks derived by Gaussian Graphical Models (GGMs) for women with and without breast cancer to better understand how foods are consumed in relation to each other according to disease status. Methods: A total of 134 women with breast cancer and 267 hospital controls were selected from referral hospitals of Tehran, Iran. Dietary intakes were evaluated by using a validated 168 food-items semi-quantitative food frequency questionnaire. GGMs were applied to log-transformed intakes of 28 food groups to construct outcome-specific food networks. Results: Among cases, a main network containing intakes of 12 central food groups (vegetables, fruits, nuts and seeds, olive oil and olive, processed meat, sweets, salt, soft drinks, fried potatoes, pickles, low-fat dairy, pizza) was detected. In controls, a main network including six central food groups (liquid oils, vegetables, fruits, sweets, fried potatoes and soft drinks) was identified. Conclusions: The findings of this study revealed a difference in GGM-identified networks graphs between cases and controls. Overall, GGM may provide additional understanding of relationships between diet and health. © 2020 Taylor & Francis Group, LLC.